16,2 Fisher: Low-Sun Phenomena in Luzon 157 
also carefully studied at various times by comparison with the 
large clock of the University physics department, which was 
also regularly compared with the time ball. I think that un- 
corrected errors in the standard time of an observation were 
never over 5 seconds. For reading the watch at night a pocket 
lens and the light of a distant street lamp were used. This 
simple equipment is precise enough for the determination of so 
unsharp a phenomenon as the appearance of a recognizable blue 
in the neighborhood of the zenith. 
While precision in the determination of the zenith is in the 
nature of the case unnecessary, nevertheless I experimented to 
see how accurately I could determine it. Thus, I faced north- 
ward and looked upward and estimated the position of the zenith 
among known stars ; then faced southward and estimated again. 
Halfway between two such estimated positions was used to com- 
pute the latitude of the observing point, by finding the declina- 
tion of the zenith on star maps in Winslow Upton’s Star Atlas. 
The results of various trials were as follows: 
(1) 
a Pegasi indistinguishably near zenith 
14 
46 
(2) 
y Pegasi indistinguishably near zenith 
14 
44 
(3) 
interpolation, Saturn and v Leonis 
15 
13 
(4) 
Interpolation, e Virginis and 42 Comae 
14 
39 
(5) 
Interpolation, a Tauri and c Tauri 
14 
18 
(6) 
Interpolation, f Tauri and Orionis 
14 
52 
(U 
/3 Leonis indistinguishably near zenith 
15 
1 
Mean declination of zenith 
14 
48 
True latitude, about 
14 
35.4 
Table 2 shows the year’s results for the zenith passage of the 
blue dawn, computed with the American Nautical Almanac and 
four-place logarithms for a point in Manila, latitude north 14° 
35.4', longitude east 8“ 3” 54^ The accuracy of tenths of minutes, 
or even minutes, is of course illusory; but an attempt to apply 
principles of the precision of measurements was not regarded 
worth the trouble. No observations were attempted with the 
moon up. Under approximately similar conditions, within a 
few days of one another results would agree within 15' or 20', 
about the sun’s semidiameter. 
Fault may be found with zenith observations of twilight, that 
the blue light may come, not from exceedingly high air directly 
illuminated by sunlight, but from air indirectly illuminated by 
light reflected from the distant atmosphere and also from high 
clouds beyond the horizon. 
